Journal of the American College of Cardiology
2014 by the American College of Cardiology Foundation
Published by Elsevier Inc.
Vol. 63, No. 20, 2014
ISSN 0735-1097/$36.00
http://dx.doi.org/10.1016/j.jacc.2014.02.549
Major Bleeding in Patients With
Atrial Fibrillation Receiving Apixaban or Warfarin
The ARISTOTLE Trial (Apixaban for Reduction in Stroke
and Other Thromboembolic Events in Atrial Fibrillation):
Predictors, Characteristics, and Clinical Outcomes
Elaine M. Hylek, MD, MPH,* Claes Held, MD, PHD,y John H. Alexander, MD, MHS,z
Renato D. Lopes, MD, PHD,z Raffaele De Caterina, MD, PHD,x Daniel M. Wojdyla, MS,z
Kurt Huber, MD,k Petr Jansky, MD,{ Philippe Gabriel Steg, MD,# Michael Hanna, MD,**
Laine Thomas, PHD,z Lars Wallentin, MD, PHD,y Christopher B. Granger, MDz
Boston, Massachusetts; Uppsala, Sweden; Durham, North Carolina; Chieti, Italy; Vienna, Austria;
Prague, Czech Republic; Paris, France; and Princeton, New Jersey
Objectives
This study sought to characterize major bleeding on the basis of the components of the major bleeding definition, to
explore major bleeding by location, to define 30-day mortality after a major bleeding event, and to identify factors
associated with major bleeding.
Background
Apixaban was shown to reduce the risk of major hemorrhage among patients with atrial fibrillation in the ARISTOTLE
(Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation) trial.
Methods
All patients who received at least 1 dose of a study drug were included. Major bleeding was defined according to the
criteria of the International Society on Thrombosis and Haemostasis. Factors associated with major hemorrhage
were identified using a multivariable Cox model.
Results
The on-treatment safety population included 18,140 patients. The rate of major hemorrhage among patients in the
apixaban group was 2.13% per year compared with 3.09% per year in the warfarin group (hazard ratio [HR] 0.69,
95% confidence interval [CI]: 0.60 to 0.80; p < 0.001). Compared with warfarin, major extracranial hemorrhage
associated with apixaban led to reduced hospitalization, medical or surgical intervention, transfusion, or change in
antithrombotic therapy. Major hemorrhage followed by mortality within 30 days occurred half as often in apixabantreated patients than in those receiving warfarin (HR 0.50, 95% CI: 0.33 to 0.74; p < 0.001). Older age, prior
hemorrhage, prior stroke or transient ischemic attack, diabetes, lower creatinine clearance, decreased hematocrit,
aspirin therapy, and nonsteroidal anti-inflammatory drugs were independently associated with an increased risk.
Conclusions
Apixaban, compared with warfarin, was associated with fewer intracranial hemorrhages, less adverse consequences
following extracranial hemorrhage, and a 50% reduction in fatal consequences at 30 days in cases of major
hemorrhage. (J Am Coll Cardiol 2014;63:2141–7) ª 2014 by the American College of Cardiology Foundation
From the *Boston University Medical Center, Boston, Massachusetts; yUppsala
Clinical Research Center, Uppsala University, Uppsala, Sweden; zDuke Clinical
Research Institute, Duke University Medical Center, Durham, North Carolina;
xG. d’Annunzio University, Chieti, Italy; kWilhelminen Hospital, Vienna, Austria;
{Motol University Hospital, Prague, Czech Republic; #Hôpital Bichat-Claude
Bernard, Paris, France; and **Bristol-Myers Squibb, Princeton, New Jersey. The
ARISTOTLE trial was supported by Bristol-Myers Squibb and Pfizer. Dr. Hylek is a
consultant with, receives travel support from, and is an adjudication committee
member with Bristol-Myers Squibb, Daiichi Sankyo, Merck, Ortho-McNeil, Johnson
& Johnson, and Pfizer; and has received lecture fees from Boehringer Ingelheim. Dr.
Held has received grants from AstraZeneca, Merck, GlaxoSmithKline, Roche, and
Bristol-Myers Squibb; is on the membership advisory board for AstraZeneca; and has
received honoraria from AstraZeneca. Dr. Alexander has received grants from BristolMyers Squibb, Merck, and Regado Biosciences; travel support from Bristol-Myers
Squibb; and consulting fees from Bristol-Myers Squibb, Pfizer, Merck, AstraZeneca,
Boehringer Ingelheim, Ortho-McNeil-Janssen Pharmaceuticals, PolyMedix, Regado
Biosciences, Bayer, and Daiichi Sankyo. Dr. Lopes has received grants from Bristol-
Myers Squibb, AstraZeneca, Boehringer Ingelheim, and Daiichi Sankyo; and
consulting fees from Bristol-Myers Squibb, Pfizer, Bayer, and Janssen Research &
Development, LLC. Dr. De Caterina is a steering committee member, a national
coordinator for Italy, coauthor of the APPRAISE-2, ARISTOTLE, AVERROES
studies, and coauthor of European Society of Cardiology Guidelines on Atrial Fibrillation;
and has received fees, honoraria, and research funding from Sanofi-Aventis, Boehringer Ingelheim, Bayer, Bristol-Myers Squibb/Pfizer, and Daiichi Sankyo. Dr. Huber
has received lecture fees from AstraZeneca, Bristol-Myers Squibb/Pfizer, Boehringer
Ingelheim, Bayer, Daiichi Sankyo, and Sanofi-Aventis. Dr. Steg has received travel
support from Bristol-Myers Squibb; is on the membership boards of Bayer, BristolMyers Squibb/Pfizer, AstraZeneca, and Boehringer Ingelheim; has received consulting fees from Bristol-Myers Squibb, Eisai, Ablynx, Amarin, Astellas, Eli Lilly,
Medtronic, Novartis, Roche, Servier, The Medicines Company, Sanofi, and AstraZeneca; grants from Servier, Sanofi, and New York University School of Medicine;
and lecture fees from Pfizer, Amgen, Otsuka, and Aterovax. Dr. Hanna is an employee
of Bristol-Myers Squibb and receives stock as part of compensation. Dr. Wallentin has
received grants from AstraZeneca, Merck, Boehringer Ingelheim, Bristol-Myers
2142
Hylek et al.
Risk Factors for Anticoagulation in Atrial Fibrillation: A Dilemma
Atrial fibrillation (AF) is a potent
risk factor for stroke. Warfarin is
highly efficacious in reducing this
AF = atrial fibrillation
risk, but its effectiveness in clinCI = confidence interval
ical practice is challenged by its
HR = hazard ratio
variable dose response, need for
INR = international
frequent monitoring, and associnormalized ratio
ated risk of hemorrhage. Among
ISTH = International Society
patients age 65 years or older,
on Thrombosis and
warfarin was noted to be the
Haemostasis
drug most often implicated in
TIA = transient ischemic
medication-related adverse events
attack
leading to emergency hospital
stay (1). Apixaban, a factor Xa inhibitor, was shown to
reduce the risk of major hemorrhage by 31% compared
with warfarin among patients with AF in the ARISTOTLE (Apixaban for Reduction in Stroke and Other
JACC Vol. 63, No. 20, 2014
May 27, 2014:2141–7
The ARISTOTLE trial design has been reported previously (3). Patients with AF and at least 1 risk factor for
stroke were randomized to receive either dose-adjusted
warfarin or apixaban, 5 mg twice daily. A reduced dose of
apixaban, 2.5 mg twice daily, was designated for participants
with 2 or more of the following criteria: age 80 years,
weight 60 kg, or serum creatinine concentration 1.5
mg/dl (133 mmol/l). The reduced dose of apixaban was
administered to 428 patients (4.7%). To enhance the
quality of warfarin management, a dosage algorithm was
provided, and a program implemented to provide regular
feedback to sites regarding their level of international
normalized ratio (INR) control.
The analyses of bleeding events included all patients who
received at least 1 dose of a study drug and included all
events from the time of the first dose until 2 days after the
last dose was received. Major bleeding was defined according
to International Society on Thrombosis and Haemostasis
(ISTH) criteria as clinically overt bleeding accompanied by a
decrease in the hemoglobin level of at least 2 g/dl or
transfusion of at least 2 units of packed red cells, occurring at
a critical site (intracranial, intraocular, intraspinal, intraarticular, intramuscular with compartment syndrome, pericardial, retroperitoneal), or resulting in death (4). No time
restrictions were applied to this definition. Laboratory and
transfusion data coupled with clinical event details were used
to identify and adjudicate potential bleeding events. Routine
collection of hemoglobin occurred every 3 months. Location
of bleeding was extracted from the case report form. Additional source documents were collected when necessary. The
primary safety outcomes were adjudicated on the basis of
pre-specified criteria by a clinical events committee whose
members were not aware of study group assignments.
Severity of hemorrhage and 30-day mortality following
first ISTH major hemorrhage. Parameters to assess the
severity of major hemorrhage, in addition to anatomic
location, for apixaban and warfarin were determined and
compared. Metrics relevant for major extracranial hemorrhage included decrease of hemoglobin of at least 2 g/dl,
hospitalization because of bleeding, transfusion of packed
red cells, number of units transfused, medical or surgical
consultation or evaluation, medical or surgical intervention
to stop the bleeding, hemodynamic compromise, and change
in antithrombotic therapy. Thirty-day mortality rates
following first ISTH major hemorrhage were evaluated and
compared between warfarin- and apixaban-treated patients.
Statistical analyses. Categorical variables were summarized
as frequencies and percentages and continuous variables as
medians and 25th and 75th percentiles. p Values representing comparisons between patients with and without
major bleeding were based on Cox regression models with
ISTH criteria first major hemorrhage as a dependent variable. p Values for the interactions between randomized
treatment and each covariate were derived using Cox
models. Factors associated with the first ISTH major
hemorrhage were identified using a multivariable Cox
model. Candidate variables included demographics and
clinical characteristics, medications, and laboratory values at
baseline. Randomized treatment and region of enrollment
were also included as candidate variables. Missing values
in predictors were imputed using multiple imputations.
Squibb/Pfizer, and GlaxoSmithKline; consulting fees from Merck & Co, Regado
Biosciences, Evolva, Portola, C.S.L. Behring, Athera Biotechnologies, Boehringer
Ingelheim, AstraZeneca, GlaxoSmithKline, and Bristol-Myers Squibb/Pfizer; lecture
fees from AstraZeneca, Boehringer Ingelheim, Bristol-Myers Squibb/Pfizer, GlaxoSmithKline, and Merck & Co; honoraria from Boehringer Ingelheim, AstraZeneca,
Bristol-Myers Squibb/Pfizer, GlaxoSmithKline, and Merck & Co; and travel support
from AstraZeneca and Bristol-Myers Squibb/Pfizer. Dr. Granger has received grants
from Boehringer Ingelheim, Bristol-Myers Squibb, GlaxoSmithKline, Medtronic
Foundation, Merck & Co, Pfizer, Sanofi-Aventis, Takeda, and The Medicines
Company; and consulting fees from Boehringer Ingelheim, Bristol-Myers Squibb,
GlaxoSmithKline, Hoffmann-La Roche, Novartis Pharmaceutical Company, Lilly,
Pfizer, Sanofi-Aventis, Takeda, The Medicines Company, and AstraZeneca. All other
authors have reported that they have no relationships relevant to the contents of this
paper to disclose.
Manuscript received August 29, 2013; revised manuscript received January 13,
2014, accepted February 16, 2014.
Abbreviations
and Acronyms
See page 2148
Thromboembolic Events in Atrial Fibrillation) trial (2).
In this report, we sought to: 1) define 30-day mortality
after a major bleeding event and determine whether this
factor differed between warfarin- and apixaban-treated
patients; 2) identify predictors of major bleeding and
determine whether predictors of major bleeding varied
between warfarin- and apixaban-treated patients; 3) further
characterize the reduction in major bleeding based on the
components of the major bleeding definition and determine
whether these components varied between warfarin- and
apixaban-treated patients; and 4) explore major bleeding
by location and determine whether bleeding locations
varied between warfarin- and apixaban-treated patients.
Methods
Baseline Characteristics
ISTH Major Bleeding ¼ Yes
Overall
(n ¼ 789)
Age, yrs
74 (68, 79)
Apixaban
(n ¼ 327)
74 (67, 79)
ISTH Major Bleeding ¼ No
Warfarin
(n ¼ 462)
74 (68, 79)
Overall
(n ¼ 17,351)
70 (62, 76)
70 (63, 76)
3,118 (35.6)
Warfarin
(n ¼ 8590)
p Value*
70 (62, 76)
<0.001
0.50
0.84
0.084
3,005 (35.0)
Interaction
p Valuey
Female (%)
270 (34.2)
102 (31.2)
168 (36.4)
Systolic BP, mm Hg
130 (120, 140)
130 (120, 140)
130 (120, 140)
130 (120, 140)
130 (120, 140)
130 (120, 140)
0.16
0.36
80 (68, 93)
83 (70, 94)
78 (67, 91)
82 (70, 96)
82 (70, 96)
82 (70, 96)
<0.001
0.012
Weight, kg
6,123 (35.3)
Apixaban
(n ¼ 8761)
Prior MI (%)
143 (18.1)
61 (18.7)
82 (17.8)
2,437 (14.1)
1,257 (14.4)
1,180 (13.8)
0.003
0.78
Prior clinically relevant or spontaneous
bleeding episodes (%)
186 (23.6)
82 (25.2)
104 (22.5)
2,847 (16.4)
1,442 (16.5)
1,405 (16.4)
<0.001
0.33
History of falls within previous year (%)
57 (7.9)
26 (8.7)
31 (7.4)
695 (4.4)
360 (4.5)
335 (4.3)
<0.001
0.57
0.078
0.75
Type of AF (%)
Paroxysmal
104 (13.2)
44 (13.5)
60 (13.0)
2,672 (15.4)
1,327 (15.1)
1,345 (15.7)
Persistent or permanent
685 (86.8)
283 (86.5)
402 (87.0)
14,676 (84.6)
7,432 (84.9)
7,244 (84.3)
JACC Vol. 63, No. 20, 2014
May 27, 2014:2141–7
Table 1
Qualifying risk factors (%)
375 (47.5)
151 (46.2)
224 (48.5)
5,280 (30.4)
2,685 (30.7)
2,595 (30.2)
<0.001
0.39
188 (23.8)
80 (24.5)
108 (23.4)
3,335 (19.2)
1,660 (18.9)
1,675 (19.5)
0.001
0.61
HF or reduced LVEF
275 (34.9)
112 (34.2)
163 (35.3)
6,160 (35.5)
3,115 (35.6)
3,045 (35.4)
0.75
0.74
Diabetes
226 (28.6)
112 (34.2)
114 (24.7)
4,300 (24.8)
2,164 (24.7)
2,136 (24.9)
0.0056
0.0034
Hypertension requiring treatment
671 (85.0)
277 (84.7)
394 (85.3)
15,188 (87.5)
7,655 (87.4)
7,533 (87.7)
0.028
0.96
Medications at randomization (%)
Amiodarone
Aspirin
Clopidogrel
82 (10.6)
31 (9.6)
51 (11.4)
1,965 (11.5)
974 (11.3)
991 (11.7)
254 (32.2)
115 (35.2)
139 (30.1)
4,159 (24.1)
2,107 (24.1)
2,052 (24.0)
0.63
0.49
<0.001
0.19
25 (3.2)
11 (3.4)
14 (3.0)
312 (1.8)
158 (1.8)
154 (1.8)
0.0026
Statin
381 (46.8)
157 (48.5)
204 (45.5)
7,081 (41.5)
3,579 (41.6)
3,502 (41.3)
0.0043
NSAIDs
100 (13.0)
40 (12.3)
60 (13.4)
1,418 (8.3)
711 (8.3)
707 (8.3)
<0.001
0.69
Gastric antacid drugs
197 (25.5)
86 (26.5)
111 (24.8)
3,139 (18.4)
1,589 (18.5)
1,550 (18.3)
<0.001
0.62
<0.001
0.042
Renal function covariates (%)
Normal (>80 ml/min)
215 (27.3)
96 (29.5)
119 (25.9)
7,281 (42.1)
3,654 (41.9)
3,627 (42.4)
Mild impairment (>50–80 ml/min)
356 (45.3)
157 (48.2)
199 (43.3)
7,209 (41.7)
3,650 (41.8)
3,559 (41.6)
Moderate impairment (>30–50 ml/min)
189 (24.1)
66 (20.2)
123 (26.7)
2,548 (14.8)
1,291 (14.8)
1,257 (14.7)
26 (3.3)
7 (2.1)
19 (4.1)
242 (1.4)
129 (1.5)
113 (1.3)
8 (1.0)
4 (1.2)
4 (0.9)
505 (2.9)
261 (3.0)
244 (2.8)
Severe impairment (30 ml/min)
Liver dysfunction (%)
0.0024
0.80
0.36
0.68
Hematocrit
42.0 (38.6, 45.0)
42.0 (38.2, 45.0)
42.0 (38.9, 45.0)
43.0 (40.0, 46.0)
43.0 (40.0, 46.0)
43.0 (40.0, 46.0)
<0.001
0.60
Platelet count, 109/l
211 (176, 253)
207 (172, 249)
213 (180, 255)
211 (178, 250)
210 (177, 249)
212 (179, 251)
0.83
0.45
Values are median (25th, 75th percentile) or n (%). *p value comparing patients with and without ISTH major bleeding. yp value for treatment by covariable interaction.
AF ¼ atrial fibrillation; BP ¼ blood pressure; HF ¼ heart failure; ISTH ¼ International Society on Thrombosis and Haemostasis; LVEF ¼ left ventricular ejection fraction; MI ¼ myocardial infarction; NSAIDs ¼ nonsteroidal anti-inflammatory drugs; SE ¼ systemic embolism;
TIA ¼ transient ischemic attack.
Hylek et al.
Risk Factors for Anticoagulation in Atrial Fibrillation: A Dilemma
Age 75 yrs
Prior stroke, TIA, or SE
2143
Hylek et al.
Risk Factors for Anticoagulation in Atrial Fibrillation: A Dilemma
2144
Missing values were uncommon (<3%) for all predictors,
except for history of fall (9%). Twenty-five imputed datasets
were generated, and a stepwise selection method was used in
each dataset. Predictors selected in more than 80% of the
imputed datasets were included in the final model. We
tested for interactions between variables in the final model
and randomized treatment.
Four additional endpoints were defined according to the
consequences of hemorrhage, including major extracranial
hemorrhage, followed by hospitalization, medical or surgical
intervention, transfusion, and change in antithrombotic
therapy. These endpoints were summarized overall and by
randomized treatment as rates (except for the number of
units of packed cells transfused), number of events, and
hazard ratios (HR) comparing all patients randomized to
apixaban versus those to warfarin. All statistical analyses
were performed using SAS version 9.2 software (SAS
Institute, Inc., Cary, North Carolina).
Results
As previously reported, the ARISTOTLE trial enrolled
18,201 patients from 1,034 clinical sites in 39 countries. The
on-treatment safety population included 18,140 patients.
The median follow-up time was 20.5 months. Major
hemorrhage occurred in 789 patients (4.3%) overall; 327 in
the apixaban group (2.13% per year) compared with 462 in
the warfarin group (3.09% per year; HR 0.69, 95% confidence interval [CI]: 0.60 to 0.80; p < 0.001). Patients who
sustained a major bleed were older (74 vs. 70 years,
respectively), more commonly had a history of myocardial
infarction, prior hemorrhage, impaired renal function, and a
fall within the previous year compared with patients without
Table 2
JACC Vol. 63, No. 20, 2014
May 27, 2014:2141–7
ISTH major hemorrhage (Table 1). They also weighed less
and had a lower hematocrit level at baseline. Among the
qualifying risk factors, patients who sustained a major
hemorrhage were more likely to have a history of stroke,
transient ischemic attack (TIA) or systemic embolism, diabetes, or hypertension. They were also more likely to use
aspirin, clopidogrel, nonsteroidal anti-inflammatory drugs,
statins, and gastric antacid drugs at baseline.
Location of hemorrhage. The most frequent sites of major
hemorrhage were gastrointestinal (31%; n ¼ 248), intracranial (22%; n ¼ 171), and soft tissue (10%; n ¼ 75)
(Table 2). Two-thirds of the gastrointestinal bleeds involved
the upper tract. Apixaban was associated with fewer
gastrointestinal hemorrhages than warfarin, but this difference did not achieve statistical significance (HR 0.89, 95%
CI: 0.70 to 1.14). There were also fewer soft tissue hematomas associated with apixaban that met the criteria for
ISTH major hemorrhage (HR 0.46, 95% CI: 0.29 to 0.74).
In addition, apixaban was associated with fewer major
hemorrhages related to trauma: 37 in the apixaban group
(0.24% per year) compared with 60 in the warfarin group
(0.40% per year; HR 0.60, 95% CI: 0.40 to 0.91;
p ¼ 0.015). As previously reported, apixaban was associated
with fewer intracranial hemorrhages than warfarin (HR
0.42, 95% CI: 0.30 to 0.58).
Severity and short-term consequences of hemorrhage.
Major extracranial hemorrhage-associated adverse consequences occurred less frequently in the apixaban group than in
the warfarin group, including fewer hospitalizations (HR:
0.75, 95% CI: 0.61 to 0.92), fewer medical or surgical interventions to stop the bleeding (HR: 0.72, 95% CI: 0.56 to
0.93), fewer transfusions (HR: 0.71, 95% CI: 0.57 to 0.89),
and fewer changes in antithrombotic therapy (HR: 0.78, 95%
Major Bleeding by Location (First Major Hemorrhage at Each Location)
Location
Intracranial
Overall
(n ¼ 18,140)
Apixaban
(n ¼ 9088)
Warfarin
(n ¼ 9052)
Apixaban vs. Warfarin
HR (95% CI)
p Value*
0.57 (174)
0.33 (52)
0.80 (122)
0.42 (0.30–0.58)
<0.001
Intra-articular
0.05 (16)
0.04 (6)
0.07 (10)
0.59 (0.21–1.61)
0.30
Digestive tract
0.83 (254)
0.78 (121)
0.88 (133)
0.89 (0.70–1.14)
0.35
Upper gastrointestinal
0.49 (151)
0.43 (66)
0.56 (85)
0.76 (0.55–1.05)
0.094
Lower gastrointestinal
0.24 (75)
0.25 (39)
0.24 (36)
1.06 (0.67–1.67)
0.80
Hemorrhoidal
0.02 (7)
0.03 (4)
0.02 (3)
1.31 (0.29–5.83)
0.73
Rectal
0.08 (26)
0.09 (14)
0.08 (12)
1.14 (0.53–2.47)
0.74
Hemoptysis
0.01 (4)
0.01 (1)
0.02 (3)
0.33 (0.03–3.15)
0.33
Hemothorax
0.02 (5)
0.01 (2)
0.02 (3)
0.66 (0.11–3.92)
0.64
Intramuscular
0.01 (2)
0.01 (1)
0.01 (1)
0.98 (0.06–15.60)
0.99
Bruising/ecchymosis
0.04 (11)
0.02 (3)
0.05 (8)
0.37 (0.10–1.38)
0.14
Epistaxis
0.07 (23)
0.08 (12)
0.07 (11)
1.06 (0.47–2.41)
0.88
Retroperitoneal
0.02 (7)
0.01 (2)
0.03 (5)
0.39 (0.08–2.02)
0.26
Intraspinal
0.01 (4)
0.01 (2)
0.01 (2)
0.98 (0.14–6.97)
0.99
Vaginal
0.02 (7)
0.03 (5)
0.01 (2)
2.46 (0.48–12.68)
0.28
Hematoma
0.25 (78)
0.16 (25)
0.35 (53)
0.46 (0.29–0.74)
0.0015
Hematuria
0.14 (42)
0.13 (20)
0.14 (22)
0.89 (0.49–1.63)
0.71
Intraocular
0.15 (47)
0.18 (28)
0.13 (19)
1.45 (0.81–2.59)
0.21
Values are rate (per 100 patient-years of follow-up) (number of events). *p values compare apixaban to warfarin rates of bleeding.
CI ¼ confidence interval; HR ¼ hazard ratio.
Hylek et al.
Risk Factors for Anticoagulation in Atrial Fibrillation: A Dilemma
JACC Vol. 63, No. 20, 2014
May 27, 2014:2141–7
CI: 0.64 to 0.95) (Table 3). Major ISTH hemorrhage criteria
followed by death within 30 days occurred half as often in the
apixaban group compared with the warfarin group, with 36
and 71 events, respectively (HR: 0.50, 95% CI: 0.33 to 0.74;
p < 0.001) (Fig. 1).
Independent factors associated with first major
hemorrhage. Older age, prior hemorrhage, prior stroke or
TIA, diabetes, lower creatinine clearance (5), and decreased
hematocrit level were independently associated with an
increased risk of major hemorrhage (Table 4). Use of aspirin
and nonsteroidal anti-inflammatory drugs also independently increased the risk of major bleeding by approximately
30%. In the multivariable models, randomization to apixaban, compared with warfarin, was associated with a lower
risk of major hemorrhage (HR: 0.69, 95% CI: 0.60 to 0.72),
as was female sex and liver disease.
In an exploratory analysis of subgroup by treatment interactions, we found a differential effect by treatment according to 3 variables: baseline renal function, weight, and
diabetes. For patients with renal dysfunction and low body
weight, the reduction in bleeding with apixaban appeared to
be greater than in patients with normal renal function and
higher body weight than in patients taking warfarin. For
patients with diabetes, the reduction in bleeding with
apixaban appeared to be less than for patients without
diabetes.
Discussion
In the ARISTOTLE trial, there were fewer intracranial
hemorrhages on apixaban, fewer adverse consequences of
extracranial hemorrhages, and fewer trauma-related hemorrhages. Apixaban, compared with warfarin, was associated
with 50% less ISTH major hemorrhage leading to death
within 30 days after the event. Warfarin treatment was
more often associated with bleeding requiring hospitalization, transfusion, procedures to stop bleeding, and change
in antithrombotic therapy, all of which may have contributed to the differences in severity and mortality. Although
the mechanisms underlying these differences are unknown,
the prolonged half-life of warfarin and warfarin’s suppression of factor VIIa may be implicated; an active factor VIIa
Table 3
Figure 1
2145
Major Bleeding Following by Death Within 30 Days
CI ¼ confidence interval; HR ¼ hazard ratio.
and tissue factor complex are necessary to initiate hemostasis (6). Although reversal of warfarin with vitamin K,
fresh frozen plasma, prothrombin complex concentrates, or
recombinant factor VIIa has been shown to reduce the INR,
the effects of these interventions on clinical outcomes are
uncertain. Delays in presentation, infusion delays, incomplete INR correction, and prothrombotic risk all undermine
the effectiveness of these agents in routine practice (7–9). In
the ARISTOTLE trial, warfarin-associated major hemorrhage also more often triggered a change in antithrombotic
therapy. Cessation of warfarin therapy was recently shown
to increase the risk of thrombosis and death among individuals who had sustained a gastrointestinal hemorrhage
(10). A better understanding of the sequelae of major hemorrhage in terms of intervention, treatment, and attendant
complications would inform its optimal management in
clinical practice.
Our findings for independent factors associated with
major hemorrhage underscore the challenge of shared risk
factors for stroke and hemorrhage among individuals with
AF, especially those who are older and have had prior stroke
and renal dysfunction (11–16). Our findings also highlight
Characteristics of Major Extracranial Hemorrhage
Overall
(n ¼ 18,140)
Apixaban
(n ¼ 9088)
Warfarin
(n ¼ 9052)
Apixaban vs. Warfarin
HR (95% CI)
Led to hospitalization
1.23 (374)
1.05 (162)
1.41 (212)
0.75 (0.61–0.92)
0.0052
Fall in hemoglobin 2 g/dl
1.25 (381)
1.06 (164)
1.44 (217)
0.74 (0.60–0.91)
0.0035
p Value*
Led to transfusion
1.06 (325)
0.89 (137)
1.25 (188)
0.71 (0.57–0.89)
0.0025
Required medical or surgical consultation
1.74 (527)
1.54 (236)
1.94 (291)
0.79 (0.67–0.94)
0.0080
Required medical or surgical intervention
to stop
0.77 (236)
0.65 (100)
0.90 (136)
0.72 (0.56–0.93)
0.012
Associated with hemodynamic compromise
0.32 (97)
0.26 (40)
0.38 (57)
0.69 (0.46–1.029)
0.069
Caused changed in antithrombotic therapy
1.31 (398)
1.14 (176)
1.47 (222)
0.78 (0.64–0.95)
0.012
Values are rate (per 100 patient-years of follow-up) (number of events). *p values compare apixaban to warfarin rates of bleeding.
Abbreviations as in Table 2.
2146
Hylek et al.
Risk Factors for Anticoagulation in Atrial Fibrillation: A Dilemma
Table 4
JACC Vol. 63, No. 20, 2014
May 27, 2014:2141–7
Baseline Factors Independently Associated With Major Hemorrhage*
Chi-Square Value
p Value
HR Represents
HR (95% CI)
Hematocrit (<45%)
Parameter
38.5
<0.001
5-U decrease (under 45)
1.38 (1.24–1.52)
Age
35.3
<0.001
10-yr increase
1.36 (1.23–1.51)
Randomized treatment
26.8
<0.001
Apixaban vs. warfarin
0.69 (0.60–0.79)
Creatinine clearance (<85 ml/min/1.73 m2)
16.9
<0.001
10-U decrease (under 85)
1.11 (1.058–1.17)
Sex
14.2
0.002
Female vs. male
0.74 (0.63–0.87)
History of bleeding
14.2
0.002
Yes vs. no
1.38 (1.17–1.63)
Aspirin at randomization
13.5
0.002
Yes vs. no
1.31 (1.14–1.52)
7.3
0.0067
Yes vs. no
1.24 (1.062–1.45)
Diabetes
NSAIDs at randomization
6.7
0.0096
Yes vs. no
1.33 (1.072–1.65)
Prior stroke/TIA/SE
5.8
0.016
Yes vs. no
1.23 (1.038–1.45)
Liver disease
5.4
0.020
Yes vs. no
0.44 (0.22–0.88)
*C-index: 0.68. Region of enrollment was included in the model.
NSAIDs ¼ nonsteroidal anti-inflammatory drugs; SE ¼ systemic embolism; TIA ¼ transient ischemic attack; other abbreviations as in Table 2.
the challenges of implementing anticoagulant therapy
among individuals with a propensity for hemorrhage, as
those individuals with a prior episode are at highest risk for
recurrence. Because of the substantial overlap in risk factors
and the major disability related to ischemic stroke, reliance
on currently available hemorrhage risk scores for decisions
regarding anticoagulant therapy is problematic. These scores
were not derived to predict intracranial hemorrhage but
rather a wide spectrum of hemorrhagic complications, most
of which do not render permanent sequelae. The inability to
account for aspirin and nonprescription nonsteroidal antiinflammatory drugs is another limitation (13,17). Use of
these tools to identify modifiable risk factors amenable to
intervention will translate into fewer hemorrhages and
improved long-term persistence with anticoagulant therapy.
The potent deleterious effects of antiplatelet drugs and
nonsteroidal anti-inflammatory agents urge caution with
concomitant use and vigilance regarding indication and
duration of therapy (18,19). Alternative analgesic medications without the attendant effects on platelets and gastric
mucosa are needed for this patient population. Ascertainment of any protective effect of gastric acid suppressants in
this setting requires randomized assessment. Similar to the
AFFIRM (Atrial Fibrillation Follow-up Investigation of
Rhythm Management) trial and the studies by Beyth et al.
(12) and Shireman et al. (20), we also found diabetes mellitus to be an independent risk factor for major hemorrhage
(15). The independent contribution of diabetes mellitus may
be obscured by renal dysfunction, given their expected high
degree of correlation. Microalbuminuria has been associated
with both intracerebral hemorrhage and hematuria (21,22).
Hypertension was not found to be an independent predictor
of major hemorrhage in ARISTOTLE, which may reflect
the degree of blood pressure control achieved among trial
participants. The lower rate of major bleeding in females has
not been previously reported and warrants further study.
Only 8 patients with liver dysfunction experienced a major
hemorrhage, precluding any definitive conclusion regarding
this patient subgroup.
Study limitations. The objective of a randomized trial is to
provide the most valid, unbiased estimate of medication
effect. For this reason, individuals with anticipated difficulty
with study protocols or with heightened risk of short
exposures (adherence, excessive risk of hemorrhage) may be
underrepresented. Thus, extrapolation of trial results to these
populations should be done cautiously. However, rates of
major hemorrhage associated with warfarin in contemporary
AF trials are 2- to 3-fold higher than the rates reported in
earlier studies, likely reflecting the older age of today’s trial
participants, the higher prevalence of chronic disease and
concomitant aspirin use, and the overall broader prescription
of anticoagulant medications in current AF populations
(23–27). Thus, we believe our trial participants are representative of most patients with AF in clinical practice.
Conclusions
Compared with warfarin, apixaban was associated with a
31% reduction in risk of first major ISTH hemorrhage.
Apixaban was associated with fewer intracranial hemorrhages, fewer adverse consequences following extracranial
hemorrhages, fewer trauma-associated hemorrhages, and a
50% reduction in fatal consequences at 30 days in case of a
major hemorrhage. Therefore, concerns for complications in
case of hemorrhage during anticoagulant treatment are fewer
during apixaban than warfarin treatment.
Reprint requests and correspondence: Dr. Elaine M. Hylek, Boston
University School of Medicine, Boston Medical Center, 801 Massachusetts Avenue, 2nd Floor Suite, Boston, Massachusetts 02118.
E-mail: [email protected].
REFERENCES
1. Budnitz DS, Lovegrove MC, Shehab N, Richards CL. Emergency
hospitalizations for adverse drug events in older Americans. N Engl J
Med 2011;365:2002–12.
2. Granger CB, Alexander JH, McMurray JJ, et al. Apixaban versus warfarin in patients with atrial fibrillation. N Engl J Med 2011;365:981–92.
JACC Vol. 63, No. 20, 2014
May 27, 2014:2141–7
3. Lopes RD, Alexander JH, Al-Khatib SM, et al. Apixaban for reduction
in stroke and other Thromboembolic events in atrial fibrillation
(ARISTOTLE) trial: design and rationale. Am Heart J 2010;159:
331–9.
4. Schulman S, Kearon C. Definition of major bleeding in clinical investigations of antihemostatic medicinal products in nonsurgical patients. J Thromb Haemost 2005;3:692–4.
5. Cockcroft DW, Gault MH. Prediction of creatinine clearance from
serum creatinine. Nephron 1976;16:31–41.
6. Mackman N. The role of tissue factor and factor VIIa in hemostasis.
Anesth Analg 2009;108:1447–52.
7. Dentali F, Marchesi C, Pierfranceschi MG, et al. Safety of prothrombin complex concentrates for rapid anticoagulation reversal of
vitamin K antagonists. Thromb Haemost 2011;106:429–38.
8. Sarode R, Matevostan K, Bhagat R, Rutherford C, Madden C,
Beshay JE. Rapid warfarin reversal: a 3-factor prothrombin complex
concentrate and recombinant factor VIIa cocktail for intracerebral
hemorrhage. J Neurosurg 2012;116:491–7.
9. Baggs JH, Patanwala AE, Williams EM, Erstad BL. Dosing of
3-factor prothrombin complex concentrate for international normalized
ratio reversal. Ann Pharmacother 2012;46:51–6.
10. Witt DM, Delate T, Garcia DA, et al. Risk of thromboembolism,
recurrent hemorrhage, and death after warfarin interruption for
gastrointestinal bleeding. Arch Intern Med 2012;17:1–8.
11. Hylek EM, Singer DE. Risk factors for intracranial hemorrhage in
outpatients taking warfarin. Ann Intern Med 1994;120:897–902.
12. Beyth RJ, Quinn LM, Landefeld CS. Prospective evaluation of an
index for predicting the risk of major bleeding in outpatients treated
with warfarin. Am J Med 1998;105:91–9.
13. Gage BF, Yan Y, Milligan PE, et al. Clinical classification schemes for
predicting haemorrhage: results from the National Registry of Atrial
Fibrillation (NRAF). Am Heart J 2006;151:713–9.
14. Lip GY, Frison L, Halperin JL, Lane D. Comparative validation of a
novel risk score for predicting bleeding risk in anticoagulated patients
with atrial fibrillation. The HAS-BLED (Hypertension, Abnormal
Renal/Liver Function, Stroke, Bleeding History or Predisposition,
Labile INR, Elderly, Drugs/Alcohol Concomitantly) score. J Am Coll
Cardiol 2011;57:173–80.
15. DiMarco JP, Flaker G, Waldo AL, et al. Factors affecting bleeding risk
during anticoagulant therapy in patients with atrial fibrillation: observations from the AFFIRM study. Am Heart J 2005;149:650–6.
Hylek et al.
Risk Factors for Anticoagulation in Atrial Fibrillation: A Dilemma
2147
16. Poli D, Antonucci E, Zanazzi M, et al. Impact of glomerular filtration
estimate on bleeding risk in very old patients treated with vitamin K
antagonists. Results of EPICA study on the behalf of FCSA (Italian
Federation of Anticoagulation Clinics). Thromb Haemost 2012;107:
1100–6.
17. Fang MC, Go AS, Chang Y, et al. A new risk scheme to predict
warfarin-associated hemorrhage. J Am Coll Cardiol 2011;58:395–401.
18. Hallas J, Dall M, Andries A, et al. Use of single and combined
antithrombotic therapy and risk of serious upper gastrointestinal
bleeding: population based case-control study. BMJ 2006;333:726.
19. Mellemkjaer L, Blot WJ, Sørensen HT, et al. Upper gastrointestinal
bleeding among users of NSAIDs: a population-based cohort study in
Denmark. Br J Clin Pharmacol 2002;53:173–81.
20. Shireman TI, Mahnken JD, Howard PA, Kresowik TF, Hou Q,
Ellerbeck EF. Development of a contemporary bleeding risk model for
elderly warfarin recipients. Chest 2006;130:1390–6.
21. Umemura T, Kawamura T, Sakakibara T, Mashita S, Hotta N,
Sobue G. Microalbuminuria is independently associated with deep or
infratentorial brain microbleeds in hypertensive adults. Am J Hypertens
2012;25:430–6.
22. Shen FC, Lee CT, Sun CK, et al. Prevalence of haematuria positively
associated with urine albumin excretion in type 2 diabetes. Diabet Med
2012;29:1178–83.
23. Levi M, Hovingh K. Bleeding complications in patients on anticoagulants who would have been disqualified for clinical trials. Thromb
Haemost 2008;100:1047–51.
24. Connolly SJ, Ezekowitz MD, Yusuf S, et al. Dabigatran versus
warfarin in patients with atrial fibrillation. N Engl J Med 2009;361:
1139–51.
25. Patel MR, Mahaffey KW, Garg J, et al. Rivaroxaban versus warfarin for
nonvalvular atrial fibrillation. N Engl J Med 2011;365:883–91.
26. Atrial Fibrillation Investigators. Risk factors for stroke and efficacy of
antithrombotic therapy in atrial fibrillation. Analysis of pooled data
from five randomized controlled trials. Arch Intern Med 1994;154:
1449–57.
27. Go AS, Hylek EM, Chang Y, et al. Anticoagulation therapy for stroke
prevention in atrial fibrillation: How well do randomized trials translate
into clinical practice? JAMA 2003;290:2685–91.
Key Words: atrial fibrillation
-
bleeding
-
factor Xa inhibitor.